My guess is that sort is taking the majority of the time in your code. If your argument list is long enough passing it in as a list can take a decent chunk of time.

use Benchmark qw(cmpthese);

@a = 1..1_000_000;

sub flat { my $array = \@_ }
sub flatcopy { my @copy = @_ } 
sub byref { my $array = $_[0] } 
sub refcopy { my @copy = @$_[0] } 

cmpthese(-1, {
    flat => sub { flat(@a) }, 
    flatcopy => sub { flatcopy(@a) }, 
    byref => sub { byref(@a) }, 
    refcopy => sub { refcopy(@a) }
});
__END__
           Rate flatcopy     flat  refcopy    byref
flatcopy 3.24/s       --     -77%     -91%     -91%
flat     14.3/s     342%       --     -59%     -61%
refcopy  35.3/s     987%     146%       --      -3%
byref    36.3/s    1020%     153%       3%       --
[download]